A Model to Predict Survival in Patients With End–Stage Liver Disease

Abstract

A recent mandate emphasizes severity of liver disease to determine priorities in allocating organs for liver transplantation and necessitates a disease severity index based on generalizable, verifiable, and easily obtained variables. The aim of the study was to examine the generalizability of a model previously created to estimate survival of patients undergoing the transjugular intrahepatic portosystemic shunt (TIPS) procedure in patient groups with a broader range of disease severity and etiology. The Model for End-Stage Liver Disease (MELD) consists of serum bilirubin and creatinine levels, International Normalized Ratio (INR) for prothrombin time, and etiology of liver disease. The model's validity was tested in 4 independent data sets, including (1) patients hospitalized for hepatic decompensation (referred to as "hospitalized" patients), (2) ambulatory patients with noncholestatic cirrhosis, (3) patients with primary biliary cirrhosis (PBC), and (4) unselected patients from the 1980s with cirrhosis (referred to as "historical" patients). In these patients, the model's ability to classify patients according to their risk of death was examined using the concordance (c)-statistic. The MELD scale performed well in predicting death within 3 months with a c-statistic of (1) 0.87 for hospitalized patients, (2) 0.80 for noncholestatic ambulatory patients, (3) 0.87 for PBC patients, and (4) 0.78 for historical cirrhotic patients. Individual complications of portal hypertension had minimal impact on the model's prediction (range of improvement in c-statistic: <.01 for spontaneous bacterial peritonitis and variceal hemorrhage to ascites: 0.01-0.03). The MELD scale is a reliable measure of mortality risk in patients with end-stage liver disease and suitable for use as a disease severity index to determine organ allocation priorities.